Electrochemical investigations of poplar, spinach, cucumber, and

Nov 1, 1992 - Felix N. Buechi, Alan M. Bond, Rachel Codd, Laila N. Huq, Hans C. ... William McFarlane , Chan Li , Takamitsu Kohzuma , Mark J. Banfield...
0 downloads 0 Views 1MB Size
Inorg. Chem. 1992, 31, 5007-5014

5007

Electrochemical Investigations of Poplar, Spinach, Cucumber, and Parsley Plastocyanins at Conventional and Microsized Carbon Electrodes Felix N. BUchi,t**Alan M. Bond,+viRachel Codd,li h i l a N. HuqJ and Hans C. Freeman’vll Department of Chemical and Analytical Sciences, Deakin University, Geelong 321 7, Australia, and Department of Inorganic Chemistry, University of Sydney, Sydney 2006, Australia Received March 20, 1992 Thermodynamic and kinetic aspects of the electrochemistry of the “blue” copper protein plastocyanin from poplar, spinach, cucumber, and parsley have been studied at both macro- and microsized carbon-disk electrodes. Reversible formal potentials EFOat 3 OC have been determined by cyclic voltammetry in the range 3.6 IpH I7.6. Above pH 6.6 the reversible potentials of all four plastocyanins are independent of pH and have limiting values of 389 f 7 mVvs NHE. Below pH 6.6the reversible potentials are functionsof pH. Poplar, spinach,and cucumberplastocyanins have similar pH dependences, but the curve for parsley plastocyanin is shifted by -0.7 pH unit toward higher pH. While the reasons for this difference and for some other details of the pH dependencesare not completelyunderstood, the fact that parsley plastocyanin behaves differently from the others is consistent with antecedent observations of redox kinetics. The electrochemical kinetic behavior of the plastocyanin/graphiteinterface is complex and is influenced by a competing protein adsorption process. Under the experimental conditions employed, only a small fraction (< 10%) of the electrode surface is active so that the conventional linear diffusion model (in which the entire surface participates) is untenable. Instead, the apparent kinetic parameters (peak separations) of the voltammetric data recorded with normal macrosized electrodes are strongly influenced by nonlinear mass transport to micron- to submicron-sizedactive sites. When nonlinear diffusion is taken into account, the experiments are consistent with very fast electron transfer between plastocyanin and the active sites of the carbon electrodes. The detection of nonlinear diffusion, combined with the detection of a voltammetric response for free Cuz+,provides a sensitive probe of the integrity of the protein; the voltammogram of the intact protein is modified in the presence of apoprotein or denatured protein. The data obtained from measurements at conventionally sized electrodes are confirmed by results obtained at carbon microelectrodes. Not only is radial diffusion found to be the predominant mechanism of mass transport (as expected from the small size of the electrodes) but also partial blocking of even these electrodes can be clearly demonstrated.

There is increasing evidence that electron transfer between a metalloprotein and the surface of an electrode mimics electron transfer between a metalloproteinand its biological redox partner, in that both processes involve a high degree of selectivity.’ In previous papers, the electrochemistry of a heme protein, cytochrome c, was reinterpreted in terms of a radial diffusion model in which the microscopic properties of the electrode/solution interface play a crucial role.z It was shown that earlier interpretations using a conventional macroscopic (linear diffusion) model led to incorrect estimates of the standard heterogeneous charge-transfer rate constant at the electrode surface.lJ An essential feature of the microscopic model is that redox protein molecules are able to recognize specific sites on an electrode as highly suitable for fast electron transfer and to reject other sites as completely unsuitable. According to this hypothesis, the electrochemistry of metalloproteins at carbon electrodes is strongly influenced by nonlinear (radial) diffusion since only part of the electrode surface is electroactive.lV2 In the case of cytochrome c the hypothesis was confirmed by studies at microelectrode^.^

We now apply the same concepts to the study of a “blue” copper protein, plastocyanin. Plastocyanin(Pc) is a component of the photosyntheticelectron transport chain in the chloroplasts of higher plants and some algae. It is the subject of three recent reviews! The molecular structure of Pc from poplar leaves, which has been extensively characterized by crystallographic structure analyses,S-’ is illustrated in Figure 1. In most higher plants the Pc molecule is a single polypeptide of 99 amino acid residues plus a copper atom. There are exceptions such as parsley Pc, where the polypeptide has only 97amino acid residues. Among the Pc’s of higher plants, it is relatively common to find some microheterogeneity (minor differencesbetween the amino acid sequencesof the Pc molecules in a single species or even in a single plant);* examples which have recently received attention occur in poplar Pc and parsley P C . ~The amino acid sequencesof the Pc’s studied in the present work are shown in Figure 2. (4) (a) Sykes, A. G. Chem. Soc. Rev. 1985, 283-315. (b) Sykes, A. G. Srruct. Bonding 1990,75,175-224. (c) Sykes, A. G. Ado. Inorg. Chem. 1991,36, 377408.

( 5 ) Colman, P. M.; Freeman, H. C.; Guss, J. M.; Murata, M.; Norris, V. A.; Ramshaw, J. A. M.; Venkatappa, M. P. Nature (London) 1978,272,

Deakin University. Present address: Paul Scherrer Institut, CH-5232 Villigen, Switzerland. Present address: Department of Chemistry, La Trobe University, Bundoora 3083, Australia. 11 University of Sydney. (1) Bond, A. M.; Hill, H. A. 0. In Metal Ions in BiologicalSyszems; Sigel. H., Sigel, A., Eds.; Marcel Dekker: New York, 1991; Vol. 27, pp 431494 (see also references cited therein). (2) (a) Armstrong, F. A.; Bond, A. M.; Hill, H. A. 0.;Psalti, I. S. M.; Zoski, C. G. J. Phys. Chem. 1989,93,6485-6493. (b) Bond, A. M.; Hill, H. A. 0.;Page, D. J.; Psalti, I. S. M.; Walton, N. J. Eur. J . Biochem. 1990, +

I

f

191. 737-742. (3) Bbchi, F.N.;Bond, A. M. J. Electroanal. Chem.InterfacialEIectrochem. 1991, 314, 191-206.

319-324. (6) G u s , J. M.; Freeman, H. C. J . Mol. Biol. 1983, 169, 521-563. (7) Guss, J. M.; Harrowell, P. R.; Murata, M.; Norris, V. A.; Freeman, H. C. J. Mol. Bi01. 1986, 192, 361-387. (8) Boulter, D.; Gleaves, J. T.; Haslett, B. G.; Peacock, D.; Jensen, U. Phytochemistry 1978, 17, 1585-1 589. (9) (a) Dimitrov, M. I.; Egorov, C. A.; Donchev, A. A.; Atanasov, B. P. FEBS Lett. 1987, 226, 17-22. (b) Dimitrov, M. I.; Donchev, A. A.; Egorov, T. A. FEBS Lett. 1990, 265, 141-145. (10) Ambler, R. P. Unpublished work cited in ref 6. ( 1 1) Rother, D.; Jansen, T.;Tyagi, A.; Herrmann, R. G. Curr. Genet. 1986, 11, 171-176. (12) Ramshaw, J. A. M.; Felton, A. A. Phytochemistry 1982, 21, 13171320.

0020-166919211331-5007$03.00/0 0 1992 American Chemical Society

Biichi et al.

5008 Inorganic Chemistry, Vol. 31, No. 24, 1992

I39

Figure 1. Structureof the poplar plastocyanin (Pc) molecule, emphasizing surface residues which have been implicated in biological or chemical electron-transfer pathways (Ca diagram with Cu atom and selected side chains, adapted from ref 6). The side chains at residues 4 2 4 5 and 5961 form an acidic patch above and below the highly conserved residue Tyr83. At the Ynorthern*(top) end of the molecule, the exposed edge of the imidazole ring of a Cu ligand, His87, is surrounded by a patch of conserved or conservatively substituted hydrophobic residues. These are denoted by black Ca atoms (Leul2, Phe35, Pro36, Leu62, Leu63, Ala65, Pro86, Ala90). The principal features of the poplar Pc structure are conserved with minor variations in the Pc’s of other species. In parsley Pc two of the residues 58-60 are missing: thus the polypeptide backbone cannot have a kink protruding into the solvent in this region, and the negative charge is reduced by 1. In poplar Pc itself, residue 45 is Ser (PcI) or Ala (PcII; see text) instead of Glu.

In the oxidized form of Pc (CuIIPc), the copper atom has a distorted tetrahedral geometry, being coordinated by the N6’ atoms of His37 and His87, the S y atom of Cys84, and the S6atom of Met92.6 The coordination of the copper atom is independent of pH.6 Reductionof the protein produces an equilibriummixture of two types of molecules: a high-pH form (CuIPc) and a protonated low-pH form (HCuIPc). In CuIPc the copper atom has the same distorted tetrahedral NNSS’ coordination as in CuIIPc, with only minor dimensional changes. In HCuIPc the copper atom has trigonal NSS’coordination; the imidazole ring of His87 is protonated, dissociated from the copper atom, and rotated by 180° about CP-CT.~The rearrangement creates a high reorganization barrier which suggeststhat the low-pH form HCuIPc should be redox-inactive.’ This hypothesis is supported by bimolecular kinetic experiments with small inorganic complexes.’4J5 In direct electrochemicalmeasurements, on the other hand, reduced Pc remains redox-active even at low pH. The apparent inconsistency is independent evidence that the equilibrium between the protonated redox-inactiveand deprotonated redox-active forms of reduced Pc is rapid on the time scale of voltammetric measurements.16 Higher plant Pc’s have a charge of about -9 at neutral P H . ~ A number of the acidic residues are highly conserved and are concentratedononesideofthemolecule(Figure1).5*6The’acidic patch” comprising residues 42-45 and 59-61 extends along the surface on both sides of another highly conserved residue, Tyr83, (13) Ambler, R. P.; Sykes, A. G. Unpublished work cited in ref 4c. (14) Sinclair-Day, J. D.; Sykes, A. G. J. Chem. Soc., Dalton Trans. 1986, 2069-2073. (15) Segal, G.S.;Sykes, A. G.J. Am. Chem. Soc. 1978, 100,45854592. (16) (a) Armstrong. F.A.; Hill, H. A. 0.;Oliver, B. N.; Whitford, D. J. Am. Chem.Soc. 1985,107,1473-1476. (b)Armstrong, F.A.Srrucr.Banding 1990, 72, 137-221 (particularly pp 180-183).

and is likely to be an electrostatic recognition site for the biological electron donor to Pc, cytochromef.17 It has been suggested that the acidic patch is responsible for electrostatic hindrance in in vitro electrochemistry.’* In the present study we have used basal-planepyrolyticgraphite electrodes to make new electrochemical measurements of the reversible formal potentials EFOfor poplar, spinach, cucumber, and parsley Pc’s at several pH values. The fact that direct electrochemistryof Pc can be achieved at carbon electrodeseither by mild acidification or by promotion with multivalent inorganic ions and complexes is known from antecedent work.16J8-20The promoters may be in solution or surface-bound. The electrochemically determined potentials EFO now reported, as well as their pH dependencesat pH 7 X cms-1 was reported for conventional linear diffusion models a t low concentrations20where electrodeblockage is minimal, and a value k? > 1cms-1 was proposed for partially active electrode surfaces.3s The resulting surface blockage parameters for poplar Pc a t pH 4.7 (related to the peak separations in Figure 7) are shown in Tables I1 and 111. The fraction of blocked electrode surface, 8,

-

(37)

7

IpA1

Sabatani, E.; Rubinstein, I.; Maoz, R.; Sagiv, J. J. Electroanal. Chem.

Interfacial Electrochem. 1987, 21 9, 365-37 1. (38) Schmidt, E.; Hitzig, J.; Titz, J.; Jiittner, K.;Lorenz, W. J. Electrochim. Acta 1986, 31, 1041-1050. (39) Sternizke, K.D.; McCreery, R. L. Anal. Chem. 1990,62, 1339-1344.

(40) Stieble, M.;Jiittner, K. J. Electroanal. Chem. Interfacial Electrochem. 1990, 290, 163-180. (41) Petersen, S.L.; Tallman, D. E. Anal. Chem. 1990, 62, 459-465. (42) Gueshi, T.; Tokuda,K.; Matsuda, H. J . Electroanal. Chem. Interfacial Electrochem. 1979, 101, 29-38. (43) Tokuda, K.; Gueshi, T.; Matsuda, H.J . Electroanal. Chem. Interfacial Electrochem. 1979, 102, 41-48.

0.5 l.OI

-0.51 i

0.5

"

"

I

"

"

I

-0.1

0.2

E vs. Ag/AgCI [VI

Figure 8. Cyclic voltammograms of poplar plastocyanin (1 10 pM) at a carbon-microdisk electrode(radiusr = 6.2 pm). Only negative potential (reduction)scans are shown. Scan rate: (a) 10 m V d ; (b) 50 m V d . Other conditions are as in Figure 6.

lies in the range 0.9 I 8 I 0.9997. Since 8 is the proportion of the surface unavailable for electron transfer to or from Pc molecules, it follows from the microscopic model that blockage of the electrode surface is a major determinant of plastocyanin electrochemistry at graphite electrodes. Tables I1 and I11 also show that the effect of varying k? in the range 2 k? 2 0.1 cms-1 is small and is not significant in relation to uncertainties in other parameters of the model (such as ro). In other words, the electrontransfer rate constant is large but the actual value cannot be determined with the present electrochemical method. Microelectrode Measurements. Experiments with a carbon microelectrode (Figure 8) provide a further demonstration that radial diffusion to partially blocked electrode surfaces makes an important contribution to Pc electrochemistry. Because such an electrode is small, diffusion at scan rates such as those used in the present study must be predominantly radial, regardless of whether the electrode is blocked or unblocked. Thus the theory predicts that a sigmoidal voltammogram should be obtained under all conditions. This is shown to be the case (Figure 8). Further, the response is seen to be independent of the scan rate over the range 10-50 m V d , as predicted for a steady-state response. Proof that the microelectrode surface is blocked or unblocked can be obtained by comparing the observed and calculated values of the limiting current. Theoretical curves for the reduction of Pc a t microdisk electrodes, calculated by digital simulation using a fast quasi-explicit finite difference method,44v45are shown in Figure 9. The observed faradaic currents are consistently smaller than theoretically expected. This fact, together with the steadystate wave shape a t all scan rates tested, unambiguously shows that the carbon microelectrode is partially blocked. The same conclusion follows from another result of the calculations (not shown here), namely that the active sites where electron transfer to Pc takes place must have a cumulative area substantially smaller than the u X 6.22 pm2 (=121 X m2) area of the carbon microelectrode. The above results are consistent with the postulate of a very large heterogeneous electron-transfer rate con~tant.2~ Within the limits of precision, the half-wave potential E112 derived from the microelectrode measurements is equal to EFO,and the Tomes criterion (E3/4-E1/4) is 170mV compared with the theoretically predicted value 55 mV for a reversible proce~s.2~ It is, however,

-

(44) Feldberg, S.W. J . Electroanal. Chem. Interfacial Electrochem. 1990, 290, 49-65. (45) Bond, A. M.; Feldberg, S.W.; Greenhill, H. B.;Mahon, P.J.; Colton, R.; Whyte, T. Anal. Chem. 1992,64, 1014-1021.

Biichi et al.

5014 Inorganic Chemistry, Vol. 31, No. 24, 1992

16.0-

12.0

-

8.0

-

4.0

-

0.0

z

0.3

0.0 E - E" [VI

-0.3

Figure 9. Digital simulations of cyclic voltammograms at a microdisk electrode (radiusr = 6.2 gm) for the reduction of an electroactivespecies. Only negative potential (reduction) scans are shown. Conditions: concentration cox = 110 g M ;diffusion coefficient D = 5.0 X lo-' cm2.s-I; heterogeneous rate constant ko = 1.Ocm-s-I; transfer coefficient a! = 0.5; scan rate (a) 10 m V d and (b) 50 m V d .

obvious that thevoltammetric data recorded with microelectrodes still suffer from considerable noise with a resulting reduction in precision . Conclusions The present study clearly demonstrates that the electrochemistry of Pc at carbon electrodes cannot be considered as a simple quasi-reversible system as previously reported and that it is decisively influenced by nonuniformity of the carbon electrode surface. The concentration-dependentpeak separations in cyclic voltammograms are symptoms of protein-electrode and/or protein-protein repulsions that are probably derived from a competing adsorption process. Such phenomena may be more common and important than has been ~upposed.~Further investigationsby surface-sensitivetechniques are now required to identify the mechanism and components responsible for the blocking of the electrode surface. A strong argument in favor of the radial diffusion hypothesis is that it leads to plausible rate constants for the electron-transfer process. If the present voltammetric results are interpreted by a conventional model assuming linear diffusion to a fully active the calculated first-order rateconstants lie in the range 10-3 2 ko 2 10-5cms-1 depending on the concentration of the

solution. A model that leads to calculated first-order rate constants that are concentration-dependent is obviously unsatisfactory. When the same experimental results are interpreted in terms of partially blocked electrode surfaces, a plausible rate constant k0 10.1cm-s-' is obtained for all the concentrations and pH values examined. One of the original objectives of this work was to obtain quantitativedata that might be used toseek relationshipsbetween the reversible potentials EFO of Pc's and the primary structures and/or charge distributions. The precision of the present measurements is inadequate to achievethis objective. It is highly desirable to develop the electrochemical methodology to achieve higher precision while using, if possible, even smaller protein specimens. For the time being, the best that can be done is to note that simple electrochemical measurements interpreted via a microscopic model lead to thermodynamic parameters that are consistent with published values derived from chemical reaction rate data. Thus we confirm that the reversible potential EFOfor parsley Pc has a pH dependencedifferent from those of the other higher plant Pc's studied. This difference must be associated with differences between the respective amino acid sequences, but the precise relationship is not yet understood. A new observation is that the EFOof poplar Pc, too, has a significantly different pH dependence. Since differences may occur among other Pc's as well (Figure 8 of ref 4c), further dissociationshould await the availability of additional and more precise data. Finally, the microscopic model implies an ability of macromolecules to discriminate among sites in an electrode surface. There is not yet any evidence that any orientation of Pc with respect to an electrode active site is more productive than others (in contrast with biological electron transfer where short-range electrostatic interactions lead to familiesof preferred orientations of Pc with respect to its redox partner"). It remains to be established whether specificity in protein-lectrode interactions involves a mechanism similar to the recognitionof redox partners in biological electron-transfer reactions. Acknowledgment. We thank Dr. P. A. Lay and Miss D. Niles for helpful comments, Mr. B. A. Fields for providing Figure 1, and Mr. D. Shi for constructing the computer-graphics model of spinach plastocyanin mentioned in the text. The work was supported by the Australian Research Council (GrantsA29030581 (A.M.B.) and A28930307 (H.C.F. and J. M. Guss)) and the Swiss National Foundation for Scientific Research. (46) Nicholson, R. S. Anal. Chem. 1965, 37, 1351-1355.